DIAGNOSTIC AND CLINICAL ASPECTS OF INVASIVE FUNGAL DISEASE AFTER ALLOGENEIC HEMATOPOIETIC STEM CELL TRANSPLANTATION

(1)

DEPARTMENT OF LABORATORY MEDICINE Karolinska Institutet, Stockholm, Sweden

DIAGNOSTIC AND CLINICAL ASPECTS OF

INVASIVE FUNGAL DISEASE AFTER ALLOGENEIC HEMATOPOIETIC STEM CELL TRANSPLANTATION

Ola Blennow

Stockholm 2014

(2)

Cover picture: Thoracic CT showing a dense infiltrate with halo sign (top), and hyphae in a lung biopsy (bottom).

All previously published papers were reproduced with permission from the publisher.

Published by Karolinska Institutet.

Printed by Åtta.45 Tryckeri AB

(3)

Institutionen för laboratoriemedicin

Diagnostic and Clinical Aspects of Invasive Fungal Disease after Allogeneic

Hematopoietic Stem Cell Transplantation

AKADEMISK AVHANDLING

som för avläggande av medicine doktorsexamen vid Karolinska Institutet offentligen försvaras i föreläsningssal R64, Karolinska Universitetsjukhuset, Huddinge

Fredagen den 17 oktober 2014, kl 09.00

av

Ola Blennow MD

Huvudhandledare:

Docent Jonas Mattsson Karolinska Institutet

Institutionen för onkologi och patologi Centrum för allogen stamcellstransplantation (CAST)

Bihandledare:

Professor Per Ljungman Karolinska Institutet

Institutionen för medicin, Huddinge Enheten för hematologi

Stockholm 2014

Fakultetsopponent:

Professor Kieren Marr

Johns Hopkins University School of Medicine Department of Infectious Diseases

Betygsnämnd:

Docent Stig Lenhoff Lunds Universitet

Avdelningen för hematologi och transfusionsmedicin

Professor Christine Wennerås Göteborgs Universitet

Avdelningen för infektionssjukdomar

Professor Bertil Christensson Lunds Universitet

Avdelningen för infektionsmedicin

(4)

(5)

ABSTRACT

Invasive fungal disease (IFD) is a major complication after allogeneic hematopoietic stem cell transplantation (HSCT). Effective prophylaxis has reduced the incidence of invasive candidiasis, but invasive mold infections (IMIs)—especially invasive aspergillosis (IA)—

continue to be an important cause of non-relapse mortality. However, there are very few data regarding IFD and IMI after HSCT in the Nordic countries. The aim of this thesis work was to investigate epidemiological, diagnostic, and clinical aspects of IFD in HSCT recipients at Karolinska University Hospital, Huddinge.

In paper I, 99 patients who received reduced-intensity conditioning (RIC) were followed with weekly fungal PCR during the first 100 days after HSCT. Patients with a positive fungal PCR result were randomized to either treatment with liposomal amphotericin B or no

treatment. We found that a single positive PCR test was not associated with IFD, irrespective of treatment. The cumulative incidence rate of proven or probable IA during the first year after transplantation was 9%, and significant risk factors in a multivariate model were grades II–IV acute-graft-versus host disease (aGVHD), cytomegalovirus- (CMV-) seronegative recipient with CMV-seropositive donor, and conditioning with alemtuzumab. In paper II, a possible influence of the intensity of the conditioning on pneumonia-related death was investigated. We found no significant differences in the cumulative incidence of pneumonia- related death between patients receiving myeloablative conditioning (MAC) and those receiving RIC: early death (< 100 days after HSCT) 2.8% vs. 2.1%, and overall death 8.2%

vs. 10.5%. Etiology could be established in 40 of 60 patients (67%) who died from pneumonia, with proven or probable IMI in 19 patients (48% of patients with established etiology, 32% of all patients with pneumonia-related death). In the multivariate analyses, grades II–IV aGVHD, CMV infection, and treatment with mesenchymal stromal cells (MSCs) were factors associated with overall pneumonia-related death. In paper III, posaconazole tissue concentrations were examined in vivo. Tissue concentrations of posaconazole were analyzed in biopsies taken at autopsy of seven patients who received posaconazole prophylaxis, and they were compared with plasma concentrations in samples taken before death. Accumulation of posaconazole was found in heart, lung, kidney, and liver tissue, while concentrations in brain were approximately equal to the concentrations in plasma. The apparent tissue accumulation in vivo is in agreement with earlier in vitro findings and may explain the low incidence of breakthrough infections seen in prophylaxis trials despite low serum concentrations. In paper IV, incidence and risk factors for IMI were retrospectively investigated in 843 patients. The cumulative incidences of proven and

probable IMI were 2.2% at day 100, 5.2% after 1 year, and 6.3% after 2 years. Factors significantly associated with a new IMI were older age (risk hazard 4.26 for 41–60 years of age and 9.0 for > 60 years of age, with 0–20 years as reference), grades II–IV aGVHD, treatment with MSCs, and transplantation with female donor to male recipient. In patients with grade II aGVHD, no IMIs were seen after onset of GVHD in 113 HSCTs performed in patients < 40 years of age, compared to 14 IMIs in 106 HSCTs (13.2%) in patients > 40 years of age (p < 0.001). Twelve of these 14 patients had signs of poor immune reconstitution

(6)

before onset of IMI. In patients with grade II aGVHD, few patients who are < 40 years of age appear to need mold-active prophylaxis, whereas in patients > 40 years of age prophylaxis is indicated if there are signs of poor immune reconstitution.

(7)

LIST OF SCIENTIFIC PAPERS

I. Blennow O, Remberger M, Klingspor L, Omazic B, Fransson K, Ljungman P, Mattsson J, Ringdén O. Randomized PCR-based therapy and risk factors for invasive fungal infection following reduced-intensity conditioning and hematopoietic SCT. Bone Marrow Transplantation 2010 Dec;45 (12):1710-8.

II. Forslow U, Blennow O, LeBlanc K, Ringden O, Gustafsson B, Mattsson J, Remberger M. Treatment with mesenchymal stromal cells is a risk factor for pneumonia-related death after allogeneic hematopoietic stem cell

transplantation. European Journal of Haematology 2012 Sep;89 (3):220-7.

III. Blennow O, Eliasson E, Pettersson T, Pohanka A, Szakos A, El-Serafi I, Hassan M, Ringdén O, Mattsson J. Posaconazole concentrations in human tissues after allogeneic stem cell transplantation. Antimicrobial Agents and Chemotherapy 2014 Aug;58 (8):4941-3.

IV. Blennow O, Remberger S, Törlén J, Ringdén O, Ljungman P, Mattsson J.

Incidence and risk factors for invasive mold infections after allogeneic stem cell transplantation. Submitted manuscript.

(8)

RELATED PUBLICATIONS NOT INCLUDED IN THE THESIS

i. Blennow O, Mattsson J, Remberger M. Pre-engraftment blood stream infection is a risk factor for acute GVHD grades II-IV. Bone Marrow Transplantation 2013 Nov;48(12):1583-4.

ii. Uhlin M, Wikell H, Sundin M, Blennow O, Maeurer M, Ringden O, Winiarski J, Ljungman P, Remberger M, Mattsson J. 2014. Risk factors for Epstein-Barr virus- related post-transplant lymphoproliferative disease after allogeneic hematopoietic stem cell transplantation. Haematologica 2014 Feb;99(2):346-52.

iii. Blennow O, Ljungman P, Sparrelid E, Mattsson J, Remberger M. 2014.

Incidence, risk factors, and outcome of bloodstream infections during the pre- engraftment phase in 521 allogeneic hematopoietic stem cell transplantations.

Transplant Infectious Diseases 2014 Feb;16(1):106-14.

iv. Blennow O, Fjaertoft G, Winiarski J, Ljungman P, Mattsson J, Remberger M.

2014. Varicella-Zoster Reactivation after Allogeneic Stem Cell Transplantation without Routine Prophylaxis-The Incidence Remains High. Biology of Blood and Marrow Transplantantation 2014 Jun 7.

(9)

LIST OF ABBREVIATIONS

aGVHD acute graft-versus-host disease

A-Mn anti-mannan antibodies

APC antigen presenting cell

BAL bronchoalveolar lavage

BG 1,3-β-glucan

BG assay assays detecting 1,3-β-glucan

CI 95% confidence interval

cGVHD chronic graft-versus-host disease

CMV cytomegalovirus

CNS central nervous system

CT computed tomography

CVC central venous catheter

EAPCRI European Aspergillus PCR Initiative

EBV Epstein-Barr virus

GM galactomannan

GM test test detecting galactomannan

GVL graft-versus-leukaemia

HCT-CI hematopoietic cell transplantation specific comorbidity index

HLA human leukocyte antigen

HR hazard ratio

HSCT In this thesis this abbreviation refers to allogeneic hematopoietic stem cell transplantation only

IA invasive aspergillosis

IFD invasive fungal disease

IMI invasive mold infection

MIC minimal inhibitory concentration

Mn mannan

MSC mesenchymal stromal cell

NK cell natural killer cell

NPV negative predictive value

(12)

PCP pneumocystis pneumonia

PCR polymerase chain reaction

PPV positive predictive value

RIC reduced-intensity conditioning regimen

RH relative hazard

VZV varicella-zoster virus

(13)

1 INTRODUCTION

In order to make it easier to read the text, allogeneic hematopoietic stem cell transplantation will be abbreviated to HSCT throughout the whole thesis. (This is not entirely correct since HSCT normally stands for all hematopoietic stem cell transplantation, i.e. both allogeneic and autologous.). Only proven and probable IFD are included in the studies mentioned in the text.

(Definitions of IFD are explained in section 2.1).

1.1 HISTORY

The concept of allogeneic hematopoietic stem cell transplantation (HSCT) was born during the animal research performed in the wake of the first atomic bomb explosions. In 1949, Jacobsen found that shielding of the spleen with lead could save mice from otherwise lethal radiation (1). A short time later, it was shown that protection could also be conferred with an intravenous infusion of bone marrow (2). Investigations regarding the potential for

autologous stem cell transplantation (i.e. from the same individual) and allogeneic stem cell transplantation (i.e. from another human being) in the treatment of leukemia were initiated, and the first transplantations in humans were performed during the latter half of the 1950s.

One of the pioneers, E. D. Thomas, later received the Nobel Prize (1990) for his achievement in this field. Disappointingly, HSCT turned out to be unsuccessful—as reported in a review of 203 transplantations carried out between 1958 and 1968, with 125 graft failures, 49 cases of severe graft-versus-host disease (GVHD), and only 11 patients achieving long-term engraftment (3). It was concluded that the clinical application had been undertaken to soon and that more research in animals was needed. During the 1960s, important knowledge was obtained about high-dose conditioning regimens that reduced the risk of graft rejection, mechanisms and prophylaxis for GVHD, and human leukocyte antigen (HLA) typing (4). In the late 1960s, clinical studies in humans could be restarted, and the results have been increasingly successful ever since. Until the end of the 1990s, however, HSCT was not suitable for older patients due to the intensive conditioning. This led to the development of reduced-intensity conditioning (RIC) regimens (i.e. conditioning that allowed hematopoiesis to recover even without transplantation), with less toxicity and shorter duration of

neutropenia, but with higher risk of relapse of the underlying disease. Different RIC protocols have been developed, and RIC is now the conditioning used in more than half of the patients in Europe and USA (5).

1.2 RATIONALE AND INDICATIONS

The rationale for performing HSCT differs between malignant and non-malignant diseases. In malignant diseases, the main reason is the graft-versus-leukemia (GVL) effect, i.e. that the new immune system will attack residual leukemia cells. In addition, HSCT permits more

(14)

intensive chemotherapy since toxicity to the bone marrow is usually the main limiting factor.

The importance of the GVL effect in humans was first described by Weiden et al. in 1979.

His group showed that in patients with grade II acute GVHD (aGVHD), relapse rate was 2.5 times less than in patients with no more than grade I aGVHD and than in patients

transplanted from an identical twin (6). The same research group later found that chronic GVHD (cGVHD) also showed a clinically important GVL effect (7). The malignant indications are dominated by hematological diseases such as acute and chronic leukemia, myelodysplastic syndrome, lymphomas, and multiple myeloma.

In the non-malignant setting, HSCT is a replacement therapy in patients with congenital or acquired deficiencies of marrow function, the immune system, or storage functions. This includes a variety of diseases such as aplastic anemia, thalassaemia, severe combined immunodeficiency, and inborn errors of metabolism. Since the rationale for HSCT in these patients is in most cases to replace dysfunctional cells and not to treat malignancy, GVL—

and thus GVHD—is not needed and should be avoided using intensive GVHD prophylaxis (8).

1.3 PROCEDURES

At least two important factors must be considered when deciding whether HSCT is indicated.

First, treatment with HSCT must have a better outcome than the alternative treatment option(s). Second, the patient must be able to withstand the procedure so that the estimated non-relapse mortality due to HSCT will be low in relation to the outcome of the disease if an alternative treatment had been chosen. When the decision has been made, HLA typing of the recipient using DNA-based typing techniques is performed and the hunt for a suitable donor can begin (9). The donor should be as well HLA-matched as possible in order to minimize the risk of severe GVHD (8, 10). An HLA-matched sibling is preferred but only exists in approximately 30% of cases. The alternatives are to use a mismatched related donor (which includes haploidentical parents or children), to find a matched unrelated donor in the large registries that have been created for this purpose, or to use stem cells from umbilical cord blood (11). Hematopoietic stem cells are still poorly characterized, but are recognized by the expression of the marker CD34 on the cell surface. The cells can be harvested from bone marrow, through leukapheresis of peripheral blood after mobilization of stem cells with granulocyte colony stimulating factor, or from umbilical cord blood. The conditioning regimen is necessary to avoid rejection of the donor cells, and to kill malignant cells if the reason for HSCT is malignant disease. At our institution, in vivo T cell depletion with anti- thymocyte globulin is administered as part of the conditioning in patients with unrelated donors and patients with non-malignant disorders in order to reduce the risk of both graft failure and GVHD (12). Other T cell inhibiting drugs must also be administered before transplantation of the stem cells to dampen the reactivity of recipient T cells in order to avoid rejection of the graft and—after transplantation—to dampen donor T cells to avoid GVHD.

The most common prophylaxis regime at our institution is cyclosporine combined with a

(15)

short course of methotrexate. Cyclosporine must be continued after the transplantation until immune tolerance has been achieved, i.e. the new T cells have learned to recognize the recipient tissues as being “self” and not to launch an attack. In the absence of treatment requiring GVHD, cyclosporine can usually be tapered over 3-6 months, but is usually given for more than one year in patients with non-malignant diseases.

The hematopoietic stem cells are infused via a central venous catheter (CVC) and they home to the marrow cavity quickly, directed by surface molecules (13). The duration of the period from cell infusion to engraftment, defined as absolute neutrophil count > 0.5 × 10⁹/L in peripheral blood, differs depending on several factors such as the conditioning regimen, the source of stem cells, the number of CD34-positive cells transplanted, and the type of donor.

Normally, patients receiving peripheral blood grafts have the shortest time to neutrophil recovery (approximately 14 days), followed by bone marrow grafts (approximately 18 days), while patients receiving umbilical cord grafts can have quite a long time to engraftment (approximately 30 days) (14, 15).

1.4 IMMUNE RECONSTITUTION

The cells that constitute innate immunity usually recover both in number and function during the first 1–2 months after transplantation, dramatically reducing the risk of bacterial and fungal infections (Figure 1) (16). The epithelial barriers are often disrupted by mucositis induced by the conditioning regimen, but recover quickly after engraftment. Neutrophil counts are rapidly normalized to pre-transplant levels, but important functions, such as chemotaxis and phagocytosis, are less efficient early on and take about two months to

recover. If glucocorticoids are administered as treatment for GVHD, recovery is delayed (17).

Monocyte counts are normalized by 1 month post-transplant. Macrophages are relatively resistant to conditioning therapies, so their number does not drop substantially, but they are gradually replaced with donor macrophages over several months (18). Most of the different dendritic cells reconstitute within 6 months, with the notable exception of follicular dendritic cells in the germinal centers of the lymph nodes and spleen (19, 20). These cells seem to recover more slowly and appear to be sparse even after 1 year. Since the follicular dendritic cells play an important role in the maturation of B cells in germinal lymph nodes, this may explain the slow reconstitution of memory B cells (19). Natural killer (NK) cell counts recover during the first weeks after transplantation. During 1–3 months after HSCT, the counts are supranormal and the NK cells kill target cells more efficiently, which may have a significant anti-leukemic effect (19).

(16)

Figure 1. Approximate immune cells counts (expressed as percentage of normal count) before and after myeloablative hematopoietic stem cell transplantation. From Bosch et al.

(16). Published with permission from Lippincott Williams & Wilkins.

The reconstitution of B and T cells takes much longer. B cell reconstitution may take up to 2 years after HSCT, especially in the presence of cGVHD. B cells are undetectable for the first two months, after which they slowly rise and reach normal levels after 1–2 years. Early on, the vast majority of B cells are naïve cells, producing IgM rather than IgG or IgA—which is reminiscent of the situation in early childhood (21, 22). After transplantation, the serum isotype immunoglobulin levels are recovered in the same sequence as they develop in young children: IgM followed by IgG1/IgG3 followed by IgG2/IgG4/IgA (23). For efficient

antibody responses, CD4+ T cells are needed, cells that are subnormal for up till 2 years after transplantation (24).

(17)

The speed and quality of immune reconstitution of T cells is of vital importance for the risk of developing viral and fungal infections after HSCT. CD8+ T cell counts reach normal levels approximately one year after transplantation, as compared to two years for CD4+ T cells, giving rise to the characteristic inverse CD4/CD8 ratio compared to healthy controls.

CD8+ cells that are antigen-primed without “help” provided by CD4+ cells may develop normally but lack memory formation and proliferative capacity (25). After transplantation, T cells are regenerated by two pathways: (1) an early thymus-independent peripheral expansion of pre-existing T cells, either mature donor T cells from the graft or residual host T cells that survive the conditioning, and (2) a later thymus-dependent pathway whereby immature thymocytes are produced in the bone marrow and transported to the thymus, where they proliferate and are subsequently released as mature naïve T cells. Because the thymus- independent pathway cannot provide new T cells, the repertoire is limited early after HSCT.

In order to have a functionally complete immune reconstitution, it is vital that new naïve T cells are generated, a process that takes from 6 months up to 2 years (26). Due to the slow T cell reconstitution and the dampened T cell activity induced by GVHD prophylaxis, all patients have a moderate risk of virus and mold infections. However, the risk is much higher in patients with factors that negatively affect the thymus function and hamper immune reconstitution—older age and GVHD being two of the most important (14, 27, 28).

1.5 GVHD AND IMMUNE RECONSTITUTION

The development of acute GVHD involves three important sequential steps or phases: (1) Activation of antigen presenting cells (APCs). Tissue damage caused by the conditioning regimen releases a storm of proinflammatory cytokines that promote activation of APCs. In addition, the conditioning causes injury to the gastrointestinal tract, allowing systemic translocation of inflammatory stimuli such as lipopolysacharides (LPS), which further enhance activation of host APCs. (2) Proliferation and differentiation of donor T cells in response to host APCs. (3) Destruction of target tissues by effector T cells. The resulting tissue damage leads to release of proinflammatory mediators, contributing to the cytokine storm that fuels aGVHD (29, 30).

There are several ways in which aGVHD affects immune reconstitution, including direct effects of the cytokine storm and bone marrow suppression caused by aGVHD-induced destruction of marrow stromal cells, and indirect side effects of the treatment given for aGVHD, usually high-dose glucocorticoids (31-33). Treatment with glucocorticoids delays reconstitution of neutrophil functions, such as chemotaxis and phagocytosis, and dampens the ability of functional neutrophils to respond to pathogens (17). A negative effect of aGVHD or its treatment has also been shown on the numbers of total and cytolytic NK cells and of dendritic cells (34). In addition, the total counts of both memory and naïve B cells are reduced, possibly due to aGVHD-induced destruction of bone marrow hematopoietic niches (31, 32, 34). Reconstitution of T cells requires a functional thymus with positive and negative selection, producing tolerant mature naïve T cells, a process that is severely impeded when

(18)

the thymus is attacked by aGVHD (26). The effect of aGVHD appears to be especially important in older patients, leading to a longstanding and sometimes irreversible negative effect on thymus function, even after successful treatment of the aGVHD (27).

Chronic GVHD most often has a later onset and has features resembling autoimmune and other immunologic disorders such as scleroderma, Sjögren syndrome, primary biliary cirrhosis, and bronchioloitis obliterans (35). The diagnosis is based on diagnostic clinical signs, in combination with exclusion of other possible diagnoses, and is graded as mild, moderate, or severe (35). In contrast to acute GVHD, the pathophysiology of chronic GVHD remains poorly understood, but results in poor T cell reconstitution (30).

1.6 INFECTIONS AFTER HSCT

The spectrum and risk of infections differs with the amount of time that has passed after HSCT. These infections are usually divided into those that occur during the pre-engraftment phase, the early post-engraftment phase (< 100 days), the mid post-engraftment phase (< 1 year), and the late post-engraftment phase (> 1 year) (Figure 2). In the later phases, GVHD is the most important risk factor for infection, and in the absence of GVHD the risk is generally low after day 100.

1.6.1 Pre-engraftment phase

Due to neutropenia, mucositis, and indwelling CVCs, bacterial infections are common during the pre-engraftment phase (Figure 2). In a recent study investigating 521 HSCTs performed in our institution between 2001 and 2008, the incidence of at least one episode of bacteremia during the pre-engraftment phase was 20% (37). In accordance with other studies, the predominant etiologies were enteric gram-negative rods and gram-positive cocci, secondary to mucositis and bacterial translocation from the gastrointestinal tract, and coagulase-negative staphylococci originating from skin and CVC (37-40). Proven bacteremia is probably only the tip of the iceberg, as almost all HSCT patients have at least one episode of neutropenic fever—the majority of which are thought to be of bacterial origin. If not treated adequately, bacteremia in neutropenic patients carries a high mortality; in the event of fever, broad- spectrum antibiotics should always be administered immediately (after blood cultures have been performed) (41). The highest mortality has been reported for bloodstream infections with gram-negative rods, and prophylaxis with quinolones during neutropenia has been much debated. Several studies have shown a reduction in episodes of fever and bacteremia with prophylaxis, but no effect on overall survival. However, a Cochrane report from 2012 (including over 40 trials with more than 5,000 hematological patients) found both significantly lower all-cause mortality and infection-related mortality with quinolone

prophylaxis (42). Another meta-analysis published earlier this year and including 1,453 stem cell recipients (407 of them allogeneic) found that bacterial prophylaxis reduced the number

(19)

Time period

Pre-

engraftment (0 to 10–30 days)

Early post-

engraftment (< 100 days)

Mid post-

engraftment (< 1 year)

Late post- engraftment (> 1 year)

Risk factors

Neutropenia Mucositis Central line

Immunsuppression (GVHD prophylaxis +/- aGVHD)

Central line

Immunsuppression (GVHD prophylaxis 6 months +/- cGVHD)

Immunsuppression (cGVHD)

Bacterial

Coagulase-negative staphylococci Enteric bacteria*

Encapsulated bacteria ( S. pneumonia, H.

influenzae) Viral

HSV

BKV (hemorr.

cystitis)

EBV

VZV If cGVHD

CMV Adenovirus

Respiratory viruses (RSV, influenza, parainfluenza) Fungal

Candida

Aspergillus (and other molds) If aGVHD or cGVHD, low if not

Pneumocystis pneumonia If cGVHD

Parasitic

Toxoplasma reactivation

High risk Moderate risk

Low risk High risk, but prophylaxis usually given

*Gram-negative rods and gram-positive cocci.

Figure 2. Timeline of infections after allogeneic stem cell transplantation. Adapted from Hammond (36).

of febrile episodes and bacteremias, but not mortality (43). However, one weakness of the meta-analysis was that the survival data in patients receiving quinolone prophylaxis were mainly derived from studies including only autologous stem cell transplantations. At our institution, prophylaxis with ciprofloxacin during neutropenia has been used routinely since 1986. The incidence of gram-negative bacteremia was low after the introduction of

prophylaxis (0.3%) but has increased significantly over the years, and was 4.0% in the period 2005–2008 (37).The risk of reactivation of herpes simplex virus (HSV) 1 and 2 is high during

(20)

the pre-engraftment period, and in most centers including ours, prophylaxis with acyclovir is given during this period. Respiratory viruses such as respiratory syncytial virus (RSV), influenza virus, parainfluenza virus, adenovirus, and metapneumovirus are important pathogens after HSCT, both during the pre-engraftment phase and later. Lower respiratory tract infections with RSV are serious complications with high mortality (44).

The risk of invasive candidiasis during neutropenia is considerable, at least after myeloablative conditioning, and prophylaxis is often given. The risk of early invasive aspergillosis (IA) (< 40 days after HSCT) has decreased during the past decade(s), with the majority of infections now occurring after engraftment, and in association with GVHD.

However, this may be changing due to more patients with risk factors for early IA (such as previous IA and active leukemia) being transplanted (45). Fungal disease after HSCT will be discussed in more detail in Chapter 2.

1.6.2 Early post engraftment phase (< 100 days)

The risk of infections with enteric bacteria almost vanishes with the resolution of

neutropenia, whereas the risk of bacteremia with coagulase-negative staphylococci is reduced but remains as long as the CVC is in place. Cytomegalovirus (CMV) reactivation is common even in the absence of aGVHD; Ljungman and colleagues found that 65% of patients had at least one CMV episode, occurring at a mean time of 26 days after HSCT (46). The

cumulative incidence of CMV disease was 1.8% after 100 days and 6.3% after one year, with the most important risk factor being grades II–IV aGVHD (46). Reactivation of Epstein-Barr virus (EBV) leading to post-transplant lymphoproliferative disease (PTLD) is an important complication with high mortality. The incidence appears to have risen during the last 10 years and the cumulative incidence at our institution was found to be over 6% in 2011–2012, with a median time of onset of 71 days (47). Respiratory viruses, including adenovirus, continue to be a threat, especially in the presence of aGVHD grade II or more (44, 48). Reactivation of varicella-zoster virus (VZV) is common without routine prophylaxis (the cumulative 2-year incidence is 22.7% at our institution), and acyclovir should be prescribed to all seropositive patients for at least one year after HSCT, and longer if ongoing immunosuppression (49-51).

Prophylaxis for pneumocystis pneumonia (PCP) is mandatory. Reactivation of toxoplasma is common in seropositive patients, and prophylaxis should be given. The risk of fungal disease is low in the absence of aGVHD.

1.6.3 Mid post engraftment phase (< 1 year)

The risk of infections after the first 100 days is coupled to development of chronic GVHD (cGVHD). In the absence of cGVHD, there is a modest risk of infections with encapsulated bacteria (Streptococcus pneumonia, Haemophilus influenzae). Prophylaxis for PCP should be given at least until GVHD prophylaxis has terminated, whereas VZV prophylaxis, as

(21)

mentioned above, should be prescribed for at least the first year after HSCT. All patients should be vaccinated against influenza.

In the setting of cGVHD, the risk of infections is determined by the severity of the cGVHD and the immunosuppressive treatment necessary, especially the glucocorticoid dose.

Vaccination is very important for protection against encapsulated bacteria, and is started 3 months after HSCT (52). However, if severe cGVHD is present, prophylaxis with antibiotics may also be considered. Reactivation of CMV and EBV is not unusual, and continued monitoring may be considered in patients with severe cGVHD. Mold infections are a threat, and prophylaxis may be adequate in certain situations as discussed in section 5.4. The risk of PCP without prophylaxis is high.

1.6.4 Late post-engraftment phase (> 1 year)

The risk of infections during this late phase depends on the cGVHD status. In the absence of cGVHD, infections are usually no problem, even if there may still be a slight increased risk of VZV reactivation and of infections with encapsulated bacteria. If cGVHD is present, the risk is, as mentioned above, determined by the severity of the cGVHD and the

immunosuppressive treatment necessary. Infections to consider are bacterial infections with encapsulated bacteria, influenza, CMV, PCP, VZV, and molds. Prophylaxis for PCP and VZV (and toxoplasma, if serpositive) should be administered for as long as

immunosuppressive drugs to treat cGVHD are prescribed. Patients with severe cGVHD requiring glucocorticoids (prednisolone) ≥ 1 mg/kg have an increased risk of developing invasive mold infections (IMIs), and mold-active prophylaxis may be indicated. This will be discussed in more detail in Chapter 2 and in section 5.4.

(22)

(23)

2 INVASIVE FUNGAL DISEASE AFTER HSCT

2.1 DEFINITIONS OF INVASIVE FUNGAL DISEASE

The lack of a standard set of definitions for fungal infections used to be a major obstacle in clinical research, making it difficult to compare different clinical trials. As a result, defining criteria were published in 2002 by the European Organization for Research and Treatment of Cancer (EORTC)/National Institute of Allergy and Infectious Diseases Mycoses (MSG) Study Group (53). These criteria were revised in 2008, and are used as the “gold standard” in clinical research involving IFD (54). The infections are classified as proven, probable, or possible. Briefly, proven invasive mold infection (IMI) requires demonstration of fungal elements in tissues, accompanied by evidence of associated tissue damage (i.e. requires biopsy specimens). A proven invasive yeast infection requires recovery of a yeast isolate by culture from a sample obtained by a sterile procedure from a normally sterile site, usually blood. To reach the level of probable fungal infection, a combination of a susceptible host, clinical signs compatible with a fungal infection, and mycological evidence must be present (Table 1). Possible infection requires only a susceptible host and clinical signs compatible with a fungal infection, the most frequent being dense, well-circumscribed lesions visible on thoracic computed tomography (CT).

2.2 EPIDEMIOLOGY 2.2.1 Candida

Invasive candidiasis after HSCT was a major problem during the 1980s, with a reported incidence of over 10% and high attributable mortality (55-57). Unlike mold infections, candida infections are endogenously derived. Since neutropenic patients lack circulating cells with phagocytosis capacity, candida translocated from the gastrointestinal tract can spread to the bloodstream, giving rise to candidemia and sometimes to disseminated infection. Thus, the most important risk factor for invasive candidiasis after HSCT is neutropenia, and the risk is highest during the pre-engraftment phase. Other important risk factors during this period are mucositis, which enhances translocation from the gastrointestinal tract, and use of broad- spectrum antibiotics, which affects the normal bacterial flora of the gut and promotes colonization of candida (58). Also, having a CVC is a well-established risk factor for candidemia, something that most HSCT recipients have for at least three months after

transplantation (58). The risk rapidly declines with the return of neutrophils and resolution of mucositis. The other period that is well known to have an increased risk is after onset of GVHD, especially if it involves the gastrointestinal tract, which is the case in more than half of all patients with aGVHD (59). The resulting mucosal ulcerations disrupt the mucosal barrier and facilitate translocation of candida, and the high doses of steroids needed for treatment of the GVHD have a negative effect on neutrophil chemotaxis and phagocytosis (60).

(24)

Table 1. Criteria for probable invasive fungal disease except for endemic mycoses according to the revised EORTC/MSG criteria (53)

Probable IFD: Requires the presence of a host factor, a clinical criterion, and a mycological criterion.

Possible IFD: Requires the presence of a host factor and a clinical criterion, but mycological criteria are absent.

Host factors

Recent history of neutropenia (< 0.5 × 10⁹ neutrophils/L for > 10 days) temporally related to the onset of fungal disease.

Receipt of an allogeneic stem cell transplant.

Prolonged use of corticosteroids (excluding patients with allergic bronchopulmonary aspergillosis) at a mean minimum dose of 0.3 mg/kg of prednisone equivalent per day for 13 weeks.

Treatment with other recognized T cell immunosuppressants, such as cyclosporine, TNF-α blockers, specific monoclonal antibodies (such as alemtuzumab), or nucleoside analogues during the past 90 days.

Inherited severe immunodeficiency (such as chronic granulomatous disease or severe combined immunodeficiency).

Clinical criteria

Lower respiratory tract fungal disease: The presence of 1 of the following 3 signs on CT:

- Dense, well-circumscribed lesions(s) with or without a halo sign - Air-crescent sign

- Cavity Tracheobronchitis:

- Tracheobronchial ulceration, nodule, pseudomembrane, plaque, or eschar seen on bronchoscopic analysis Sinonasal infection: Imaging showing sinusitis plus at least 1 of the following 3 signs:

- Acute localized pain (including pain radiating to the eye) - Nasal ulcer with black eschar

- Extension from the paranasal sinus across bony barriers, including into the orbit CNS infection: 1 of the following 2 signs:

- Focal lesions on imaging

- Meningeal enhancement on MRI or CT

Disseminated candidiasis³: At least 1 of the following 2 entities after an episode of candidemia within the previous 2 weeks:

- Small, target-like abscesses (bull’s-eye lesions) in liver or spleen

- Progressive retinal exudates on ophthalmologic examination

Mycological criteria

Direct test (cytology, direct microscopy, or culture)

Mold in sputum, bronchoalveolar lavage fluid, bronchial brush, or sinus aspirate samples, indicated by 1 of the following:

- Presence of fungal elements indicating a mold

- Recovery by culture of a mold (e.g., Aspergillus, Fusarium, Zygomycetes, or Scedosporium species) Indirect tests (detection of antigen or cell-wall constituents):

- Aspergillosis: galactomannan antigen detected in plasma, serum, bronchoalveolar lavage, or CSF - Invasive fungal disease other than cryptococcosis or zygomycoses: 1,3-β-glucan detected in serum

(25)

Today, invasive candida infections after HSCT are a minor problem; large studies published during the last years have shown incidences under 2% (37, 45, 61). The main reason is probably the routine use of systemic antifungal prophylaxis during the pre-engraftment phase and during severe GVHD. Another important explanation is likely to be the increasing use of RIC and peripheral blood stem cells as stem cell source with less tissue damage, i.e.

mucositis, and shorter duration of neutropenia.

2.2.2 Aspergillus

Invasive aspergillosis (IA) after HSCT remains a major complication with high mortality. As mentioned in section 1.6.1, the timing of IA changed during the 1990s, from early onset to late onset, often occurring in association with GVHD (62-64). However, this may be

changing again as reported in a recent prospective surveillance study including 1,858 HSCTs performed in Italy in the period 2008–2010. The overall cumulative 12-month incidence of IFD in the study was 8.8%, 81% of which was due to IA. The majority of IA (57.3%) occurred early, which was thought to be due to more patients with risk factors for early infection (such as previous IA and active leukemia) being eligible for transplantation (45). A similar overall incidence of IFD was found in 6,666 HSCT recipients transplanted in the USA in 2001–2005, with 42% being due to IA (65). Numerous risk factors for development of IA after HSCT have been reported, including age > 40 years at transplantation, active leukemia at transplantation, cord blood as stem cell source, transplantation with an unrelated donor or mismatched related donor (as opposed to a matched related donor), delayed neutrophil engraftment, delayed lymphocyte engraftment, grades II–IV aGVHD, severe chronic GVHD, aGVHD before severe chronic GVHD, aGVHD in a patient transplanted with a donor other than a matched related one, and treatment with glucocorticoids (63, 66, 67). Most of these risk factors reflect poor T cell reconstitution, highlighting the importance of well-functioning T cells for avoidance of IMI. The fact that grades II–IV aGVHD is an important risk factor for IMI after HSCT has been well known for many years, but data showing the risk according to the level of aGVHD have been limited until recently. In the large study from Italy

mentioned above, the 6-month cumulative incidence of IFD in patients with grade II aGVHD was 7.1% as compared to 12.3% in patients with grades III–IV (45). Furthermore, the risk of IFD in patients with grades II– IV aGVHD varied according to whether or not aGVHD was followed by cGVHD, and to the type of donor. The cumulative incidence of IFD in patients with aGVHD not followed by cGVHD was as low as 2.3% in patients with matched related donors, as compared to 10% in patients with other donors. In patients with aGVHD followed by cGVHD, the corresponding percentages were 10% and 25.3%.

The mortality from IA reported in larger studies has varied from 35% to over 70%, partly depending on differences in follow-up time, autopsy frequency, and whether overall or attributable mortality was reported (45, 64, 65, 68-70). Survival has increased during the last decade, at least partly due to increased use of voriconazole (64). Many risk factors for death have been reported and they include impairment in pulmonary function before

(26)

transplantation, neutropenia at diagnosis of IA, early IA, late IA, treatment with corticosteroids (prednisolone) at ≥ 2 mg/kg per day, mismatched donor, proven IA (as opposed to probable), and renal insufficiency (64, 69, 70).

2.2.3 Other molds

Invasive non-Aspergillus mold infections are uncommon, but the incidence is increasing and such infections are often fatal. The most common of these uncommon infections is

mucormycosis, formerly known as zygomycosis. The reason for the change of name is a new classification of fungi together with the fact that the majority of zygomycosis infections are caused by members of the Mucorales, such as the genera Rhizopus, Mucor, Lichtheimia, Cunninghamella, and Rhizomucor (71, 72). The Mucorales can usually be found in decaying organic matter and infections are either caused by inhalation of spores or direct inoculation of wounds after penetrating trauma (73). Important risk factors after HSCT are neutropenia and receiving glucocorticoids as treatment of GVHD, whereas diabetes is the most important risk factor in non-HSCT patients (73). In a prospective surveillance study from the TRANSNET database including 16,200 stem cell transplantations performed in the USA in 2001–2006 (79% HSCT); 8% of all IFD was caused by mucormycosis, with a 12-month cumulative incidence of 0.29% (65, 72). In a study including 929 reported cases of mucormycosis, the mortality was around 40% for all patients but as high as 91% in HSCT recipients (74). In a recent study investigating risk factors for early death from pulmonary mucormycosis in hematology patients (39% HSCT recipients), the 4-week mortality was 37% for all patients but 75% in patients with high APACHE score, lymphocyte count under 0.1 × 10⁹/L, and elevated lactate dehydrogenase (75).

Another uncommon but important mold infection is fusariosis, constituting 25% of the non- Aspergillus mold infections in the TRANSNET study (72). In a recent retrospective study of 233 cases diagnosed between 1985 and 2011, 92% of the patients had a hematological

disease, 38% had undergone HSCT, and 75% were neutropenic at diagnosis (76). The 90-day survival probability improved significantly during the study period, from 22% in the period 1985–2000 to 43% in the period 2001–2011, and was highest for patients receiving

voriconazole (60%). Variables associated with poor prognosis were receipt of

glucocorticoids, neutropenia at the end of treatment, and treatment with amphotericin B deoxycholate (76).

2.3 ANTIFUNGAL AGENTS FOR SYSTEMIC USE 2.3.1 Amphotericin B derivates

Amphotericin B is the oldest antifungal drug still in regular use, with initial reports of antifungal activity published as early as 1956. Amphotericin B binds to ergosterole in the

(27)

fungal membrane, leading to alteration in permeability and leakage of cell components, and ultimately to cell death (Figure 3) (77). Only intravenous formulations can be used for systemic treatment since orally administered amphotericin B is not absorbed. The first drug available for systemic use, amphotericin B deoxycholate, had dose-limiting nephrotoxicity.

As a result, during the second half of the 1980s, lipid formulations were developed that contained amphotericin B within liposomes—biodegradable vesicles that consist of an aqueous environment surrounded by phospholipid bilayers (78). These formulations have significantly less renal toxicity and have replaced amphotericin B deoxycholate in most transplant facilities, including our own. Amphotericin B formulations have a broad antifungal activity, including Candida spp. and the majority of molds involved in IFD after HSCT, such as Aspergillus spp. (with the exception of A. terreus which has higher minimal inhibitory concentrations (MICs)), Mucorales, and Fusarium spp. (79). Liposomal amphotericin B is the only lipid compound currently available in Sweden, and the usual dose used in the treatment of mold infections is 3–5 mg/kg per day.

Figure 3. Modes of action of antifungal drugs.

Inhibit the enzyme 1,3-β-glucan synthetase and stop production of 1,3-β-glucan

Echinocandins

Amphotericin B

Binds to ergosterole and changes cell membrane permeability

Inhibit 14α-lanesterol demethylation and stop ergosterol synthesis

Azoles

Ergosterole

(28)

2.3.2 Azoles

The first azoles in clinical use, the imidazoles, were available at the beginning of the 1970s.

Miconazole was the first azole that could be used for systemic treatment, but it had to be administered intravenously and had many side effects such as phlebitis, fever, and chills (80).

Ketoconazole, introduced in 1977, was the first azole that could be used orally, but its use was limited by variability in absorption, gastrointestinal intolerance, and liver toxicity (81).

During the 1980s, a new group of azoles was synthesized, the triazoles, which all inhibit 14α- lanesterol de-methylation, a necessary step in ergosterol synthesis in the fungal cell

membrane (Figure 3) (82). Fluconazole was developed during the latter half of the 1980s, and was shown to be a well-tolerated drug with good oral bioavailability and efficacy against Candida species with the exception of C. krusei and C. glabrata (83, 84). Itraconazole was introduced at approximately the same time as fluconazole, and was the first azole with anti- mold activity. Itraconazole has been a widely used drug, especially for prophylaxis, but variability in absorption, gastrointestinal tolerance, and hepatotoxicity have limited its

usefulness. The next azole with anti-mold activity was voriconazole; it was approved in 2002 as the first of the second generation of triazoles. Voriconazole has good activity against Candida spp., Aspergillus spp, and Fusarium spp (85). Posaconazole was introduced in 2006 and has a broad antifungal activity covering Candida spp., Aspergillus spp., and some members of the Mucorales (85). There are several new azoles in clinical trials: isavuconazol, ravuconazole, and albaconazole.

2.3.3 Echinocandins

The echinocandins differ from the other antifungal agents in that they attack the cell wall and not the cell membrane. They inhibit the enzyme 1,3-β-glucan synthetase, which is necessary for production of 1,3-β-glucan (BG), a major component of the fungal cell wall which is fundamental for resistance to osmotic pressure (Figure 3). Generally, the echinocandins are well tolerated with few side effects, but they have the disadvantage of only being available as intravenous formulations. They are the first-hand choice when treating Candida infections (except C. parapsilosis and C. guilliermondii, which may have high MICs), and they are active against Aspergillus spp. but not against other molds (86-88).

At the moment, there are three registered echinocandins: caspofungin (the first to be

approved in 2001), micafungin, and andulafungin. As all echinocandins have the same target, they have a very similar spectrum of activity. Caspofungin has the best documentation for use in patients with neutropenia, whereas anidulafungin appears to have the least hepatotoxicity, but these differences are of minor importance in clinical practice (86, 89, 90).

(29)

2.4 PROPHYLAXIS

2.4.1 Pre-engraftment until day 100

The antifungal drugs available for systemic treatment were very limited during the 1980s and consisted mainly of amphotericin B deoxycholate, with dose-limiting nephrotoxicity, and ketoconazole, with variable oral absorption and risk of severe liver toxicity (81). The situation changed with the introduction of fluconazole, a well-tolerated drug with good oral bioavailability and efficacy against most Candida species (83, 84). Since the risk of invasive candidiasis is highest during neutropenia, Goodman and colleagues performed a randomized trial comparing prophylaxis with fluconazole (400 mg once daily) with placebo during the pre-engraftment phase in 356 transplant patients (48% allogeneic) (57). The results showed a significantly lower incidence of invasive candidiasis in patients receiving fluconazole (2.8%

vs. 15.8%) and lower candida-associated mortality (0.6% vs. 5.6%). In 1995, Slavin et al.

reported data from a randomized study (with a long-term follow-up published in 2000) comparing prophylaxis with fluconazole (400 mg once daily) or placebo for 75 days in 300 patients undergoing HSCT (60, 91). Again, fluconazole recipients had a significantly lower overall incidence of invasive candidiasis (2.6% vs. 20.3%) and, importantly, also

significantly better survival at 8-year follow-up (44.7% vs. 27.7%). Curiously, fluconazole prophylaxis had a prolonged protective effect: more patients in the placebo arm died of late (defined as > 110 days after HSCT) invasive candidiasis than those in the fluconazole arm (8 of 96 as opposed to 1 of 121). The explanation appeared to be that the incidence of severe GVHD involving the gut was significantly lower in fluconazole recipients, resulting in a lower risk of disseminated candida infection (60).

When an effective candida prophylaxis was used, it became evident that mold infections were major complications after HSCT, with high mortality. Subsequent prophylaxis trials therefore focused on comparing fluconazole with drugs that had activity against molds. Itraconazole prophylaxis for 180 days, or until 4 weeks after discontinuation of GVHD therapy, was compared with fluconazole, and was shown to provide significantly better protection against IMIs while patients were on treatment. However, there were no significant differences in IMI or invasive candidiasis at the end of the trial because of significantly more discontinuation of the study drug due to gastrointestinal intolerance in itraconazole recipients (92). Another, smaller study using prophylaxis for 100 days after HSCT also found itraconazole prophylaxis to be superior to fluconazole prophylaxis but, again, there were significantly more

gastrointestinal side effects (93). It was concluded that even though itraconazole is more effective in preventing fungal infections, i.e. mold infections, hepatotoxicity and

gastrointestinal intolerance limit its use.

Voriconazole, which has better gastrointestinal tolerance than itraconazole, was compared with fluconazole in a large prophylaxis study including 600 patients (61). The drugs were administered for 100 days after HSCT, or for 180 days in higher-risk patients (including aGVHD treated with prednisolone at > 1 mg/kg per day). Disappointingly, even though there was a trend of fewer IFDs in voriconazole recipients (7.3% vs. 11.2%), it did not reach

(30)

statistical significance. In addition, there was no significant difference in fungal-free survival.

The authors stated that possible explanations included intensive monitoring with galactomannan (GM) tests and structured empiric antifungal therapy, enabling early

treatment in patients with IFD and possibly reducing mortality in fluconazole recipients (61).

A subgroup analyses showed that transplantation because of acute myeloid leukemia was a risk factor for IFD and that voriconazole was protective in this group. Thus, it appeared that the risk of mold infections in a mixed cohort of HSCT recipients might be too low for mold prophylaxis to be warranted, and that such prophylaxis should be reserved for higher-risk patients. This notion was indirectly supported by a study published in 2011 that compared voriconazole and itraconazole prophylaxis for 100 days after HSCT in 489 patients, which found incidences of IFD of 1.3% and 2.1%, respectively (94). A limitation in these trials was the reduced performance of the GM test in patients who received mold prophylaxis (95). This may have led to underestimation of IA, since the GM test is included in the EORTC/MSG definitions and—in most studies—is the most common mycological criterion fulfilled. With a false-negative GM test, many IAs will be classified as possible instead of probable, and thus be excluded from analysis (54).

Posaconazole prophylaxis during the pre-engraftment has not been properly evaluated, only two small studies involving 55 posaconazole recipients have been published (96, 97).

Micafungin was compared to fluconazole during neutropenia in a large study including 882 patients undergoing allogeneic transplantation (54%) or autologous transplantation (46%).

The IFD incidence was low in both treatment arms (2.4% vs. 1.6%) and not significantly different (98). This was not unexpected, since patients undergoing autologous transplantation are not at risk of developing mold infections, so in reality the trial was underpowered.

The lipid formulations of amphotericin B have long half-lives, which may make them

suitable for administration on alternate days in the prophylactic setting (99). The small studies that have been published so far have used either a high dose once a week, with reports of nephrotoxicity, or lower doses two or three times a week (97, 100). The results from a large placebo-controlled trial using 5 mg/kg liposomal amphotericin B twice weekly in patients with acute lymphatic leukemia are pending (NCT01259713), and may give a hint of whether this approach could also be useful after HSCT.

Inhalation of liposomal amphotericin B has─compared to fluconazole─been shown to significantly decrease invasive pulmonary aspergillosis in neutropenic non-HSCT patients (101). Aerosolized formulations of amphotericin B have also been tried after HSCT and appear to be safe, but the studies have been too small to be able to draw any conclusions regarding efficacy (102). It is important to remember that this prophylaxis only protects against lung aspergillosis, and that it must be combined with fluconazole or some other candida prophylaxis.

(31)

2.4.2 GVHD

In the setting of GVHD, prophylaxis with fluconazole was shown to be effective in preventing invasive candidiasis and candida-related death in the study by Marr and

colleagues (60). Posaconazole has been shown to be as effective as fluconazole in prevention of all IFD in patients with grades II to IV aGVHD or chronic extensive GVHD, and superior in preventing IA (2.3% vs. 7.0%) at the end of the fixed 112-day treatment period (103). In the study by Wingard et al. comparing voriconazole with fluconazole, administration of the study drug was prolonged to 180 days after HSCT in high-risk patients, including patients treated with > 1 mg/kg prednisone per day on days 90–100. As mentioned above, there was a trend of fewer IFDs in voriconazole recipients (7.3% vs. 11.2%), but this did not reach statistical significance (61).

2.4.3 Secondary prophylaxis

With the introduction of RIC and less toxic antifungals (lipid formulation of amphotericin B, voriconazole and caspofungin), more patients with a history of IFD became eligible for HSCT. Retrospective studies strongly suggested a protective effect on relapse rates of IFD with secondary prophylaxis (104). A prospective study, conducted with voriconazole in 45 patients with a history of IFD (31 IA) prior to transplantation, confirmed the protective effect of secondary prophylaxis with only 3 patients developing an IFD the first year after HSCT (105). Recently, a prospective multicenter study from China reported a 1-year cumulative relapse rate of 25% in 130 patients with a history of IA before transplantation (106). In this study the antifungal used for treatment of IA was also used as secondary prophylaxis.

2.4.4 Other aspects of prophylaxis

Since the beginning of the 2000s, increasing azole resistance in clinical Aspergillus spp.

isolates has been reported. The vast majority has been due to infection with an already resistant environmental isolate, resistance that is thought to be secondary to the use of agricultural fungicides (107). Acquired resistance still appears to be rare but has been reported in patients receiving long-term treatment because of chronic forms of aspergillosis (108). Development of resistance can also be due to selection of intrinsically resistant species, as reported in a prospective multicenter study of candidemia in France, in which recent exposure of fluconazole was associated with infection with an isolate with decreased

susceptibility to fluconazole (109). Acquired azole resistance appears to be uncommon, but in a report from two tertiary cancer centers in the USA by Oxman et al., acquired rather than intrinsic resistance was thought to be the explanation in a third of cases with candidemia due to fluconazole-resistant isolates (110).

The rationale for recommending therapeutic drug monitoring when using voriconazole and posaconazole is the established dose-response effect combined with variable drug

(32)

concentrations when using standard dosages (111). The concept is uncontroversial for voriconazole because of its extremely variable pharmacokinetics. However, the

recommendation when using posaconazole appears to more due to variable absorption of the oral solution than the pharmacokinetics, which is quite stable. With the introduction of a solid tablet formulation (already available in the USA, will be available during the fall in Sweden), the problem of variable absorption appears to be solved and drug monitoring during

prophylaxis will be redundant (112).

2.4.5 Summary

Candida prophylaxis has been found to be beneficial both when administered during

neutropenia and after engraftment up to day 75. Fluconazole prophylaxis is still a valid choice during the pre-engraftment phase unless there is a high local incidence of C. glabrata or C.

krusei. The shorter duration of neutropenia after RIC probably reduces the risk of

candidemia, but the data have not been altogether convincing, and administration of a safe, low-priced drug such as fluconazole seems reasonable. The effect on late-onset candida seen in the prolonged fluconazole prophylaxis trial by Slavin and Marr is probably less important today, since most patients with severe GVHD will receive mold-active prophylaxis.

However, mold-active prophylaxis does not appear to be necessary as a routine prophylaxis after HSCT, due to the low incidence of early IMI, and should be reserved for high-risk patients such as those receiving corticosteroids because of GVHD. This will be explored in more detail section 5.4. Secondary prophylaxis is mandatory in patients with a history of IFD before transplantation. Although voriconazole has the best documented effect, it can also be an option to continue with the antifungal used for initial treatment.

2.5 DIAGNOSTICS 2.5.1 Candida

2.5.1.1 Candidemia and acute invasive candidiasis

The most common form of invasive candida infection in HSCT patients is candidemia.

Usually the only symptom is fever, sometimes combined with signs of septic embolus to skin, lungs, or eyes. Undiagnosed candidemia during neutropenia can disseminate to the liver and spleen, resulting in hepato-splenic candidiasis (also known as chronic disseminated

candidiasis), which usually cannot be diagnosed before recovery of neutrophils and the subsequent formation of multiple small abscesses.

Invasive candida infections can be quite difficult to diagnose. Blood culture is the gold standard, but the sensitivity has been reported to be only 50–75% (113). As a result, alternative tests based on the detection of fungal biomarkers and metabolites have been developed. Mannan (Mn) is a major component of the cell wall, and is one the main candida

(33)

antigens that circulate during infections (114). Initial observations showed an inverse

relationship between circulating Mn and anti-mannan antibodies (A-Mn), which can both be measured with ELISA tests (Platelia® Ag Plus and Platelia® Ab Plus; Bio-Rad laboratories).

A systematic review of studies evaluating Mn and A-Mn, published in 2010, found a

sensitivity of 83% and a specificity of 86% when combining the two tests (115). Even though there was significant heterogeneity between the studies, it was concluded that a combined test was useful for diagnosis of invasive candidiasis (115, 116). A more recent case-control study including 56 patients with candidemia and 200 controls found a similar sensitivity—89%—

but a specificity of only 63% when combining Mn and A-Mn (117). The reason for the poor specificity was the A-Mn test, which had a specificity of only 65% when used alone.

However, using Mn test alone did not appear to be an option; even though the specificity was excellent at 97.5%, the sensitivity was only 59%. One explanation for this low sensitivity was that C. parapsilosis was not detected with the test (117).

Another test targets 1,3-β-glucan (BG), a major component in the cell wall of Candida spp, Aspergillus spp, Pneumocystis jirovecii, and many other fungi with the notable exceptions of the Mucorales and Cryptococcus spp. BG is released in blood and tissues during the course of an IFD, and is usually detected in plasma or serum. Detection in cerebrospinal fluid has also been shown to be useful, whereas the performance of BG assay in bronchoalveolar lavage fluid (BAL) is reported to be poor for fungal infections other than PCP (118, 119). There are four commercial tests available, three from Japan and one from the USA (Fungitell®; Beacon Diagnostics Laboratory, Cape Cod Inc.). They are all based on amoebocyte lysate from the horseshoe crab, but use different species. In the presence of BG, a coagulation cascade is initiated in the lysate, which leads to quantifiable transformation of a chromogenic substrate.

Several prospective cohort studies have evaluated the performance of the BG assays, but differences in both populations and end-points make comparisons difficult. In addition, the cut-off values and number of positive tests required (one or two) have varied (116, 120-122).

In a meta-analysis published in 2011, the performance of the BG assays was analyzed in patients with hematological malignancies (122). Six cohort studies including 1,771 adults with 215 IFDs were analyzed. Excellent specificity was found (98.9%)—but low sensitivity (49.6%)—when 2 consecutive positive tests were required. With an IFD prevalence of 10%

(pre-test likelihood), the estimated positive and negative predictive values (PPV and NPV) were calculated to be 83.5% and 94.6%, respectively. With an IFD incidence of 2%, the PPV would decrease to 48% with a slight increase in NPV to 98.9%. Another large meta-analysis also including non-neutropenic patients, albeit with significant heterogeneity, found a pooled sensitivity of 76.8% and specificity of 85.3% in 2,979 patients with 594 proven or probable IFD (123). No difference in performance regarding Candida and Aspergillus infections was noted. Finally, another meta-analysis has shown BG assays to be of excellent value in diagnosing PCP, with a sensitivity of 94.8% (124).

The merits and usefulness of polymerase chain reaction (PCR) tests in blood specimens after HSCT will be discussed in section 5.1.

(34)

2.5.1.2 Chronic disseminated candidiasis

Reaching a diagnosis of chronic disseminated candidiasis can be difficult, since blood cultures are usually negative. BG assays and combined Mn and A-Mn testing is

recommended. Imaging of liver and spleen should be performed, looking for focal lesions, and, if found, biopsied for microscopy, culture, and PCR. Biopsies are often microscopy- negative and culture-negative, possibly due to ongoing antifungal treatment or uneven spread of yeast in the lesions, in which case PCR has been shown to be of value (125, 126).

2.5.1.3 Summary

Due to the low incidence of invasive candida infections after HSCT when using candida- active prophylaxis, non-specific use of Mn and A-Mn, BG assays, and PCR tests is discouraged. In the event of prolonged neutropenic fever, the tests may be beneficiary but since the incidence of invasive candidiasis will also be low in this group, low PPV is a problem and may lead to a false-positive diagnosis. If signs of chronic disseminated infection are present on imaging of liver or spleen, low specificity will be less of a problem and BG and combined Mn and A-Mn testing in serum is recommended. In addition to the mandatory blood cultures, biopsies should be taken for microscopy, culture, and PCR.

2.5.2 Aspergillus

An invasive Aspergillus infection starts with inhalation of conidia, which germinate to form hyphae in the sinuses or lungs. After penetration of the epithelial barrier, it becomes an angioinvasive infection with risk of local bleeding and thrombosis. Occasionally the infection spreads hematogenously and becomes disseminated. The clinical picture varies according to the time period after HSCT. Early after HSCT, during the pre-engraftment phase, the first symptom of mold infections is usually fever, often followed by cough—sometimes with hemoptysis. Occasionally the first symptoms come from organs other than lungs and sinuses, such as the brain, due to unrecognized disseminated disease. In those cases, asymptomatic lung infiltrates are usually found on thoracic computed tomography (CT). Aspergillus infections after engraftment usually occur during treatment of GVHD, and, due to the presence of neutrophils, tend to have a slower course with low-grade fever and progressive respiratory symptoms.

The diagnostic tools of importance are culture and microscopy (of sputum, BAL, and biopsies), thoracic CT, GM test (serum and BAL), PCR test (blood specimens, BAL, and biopsies), and BG assays (serum).

(35)

2.5.2.1 Culture and microscopy

Microscopy is fast, easy to perform, and helpful to establish a diagnosis when positive (Figure 4). Since molds can be difficult to get to grow, microscopy should always be performed on biopsies, and it is also useful for sputum and BAL. The disadvantages of microscopy include variable sensitivity and not reaching a species identification (127).

Figure 4a. Rhizopus microsporus in a biopsy from sinus maxillaris, hematoxylin-eosin staining (top) and fluorescent staining (bottom).